Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2014-07-16

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:25 aimon joined #puppet-openstack
00:45 rmoe joined #puppet-openstack
00:59 aldavud joined #puppet-openstack
01:05 rstarmer joined #puppet-openstack
01:08 mmaglana joined #puppet-openstack
01:15 xarses joined #puppet-openstack
01:24 gildub joined #puppet-openstack
01:49 ilbot3 joined #puppet-openstack
01:49 Topic for #puppet-openstack is now Place to collaborate on Puppet/OpenStack tools: logs at http://irclog.perlgeek.de/puppet-openstack/today
02:10 thumpba joined #puppet-openstack
02:18 xingchao joined #puppet-openstack
02:21 JayJ_ joined #puppet-openstack
02:53 aldavud joined #puppet-openstack
03:00 daneyon joined #puppet-openstack
03:01 daneyon joined #puppet-openstack
03:11 thumpba joined #puppet-openstack
03:38 saju_m joined #puppet-openstack
04:31 sanjayu joined #puppet-openstack
05:03 denca joined #puppet-openstack
05:32 rstarmer joined #puppet-openstack
06:28 ddmitriev joined #puppet-openstack
06:32 tfz joined #puppet-openstack
06:39 EmilienM good morning
07:14 artem_panchenko joined #puppet-openstack
07:19 arnaud_orange joined #puppet-openstack
07:32 aldavud joined #puppet-openstack
07:46 marcel joined #puppet-openstack
08:10 derekh_ joined #puppet-openstack
08:23 mmagr joined #puppet-openstack
08:36 qba73 joined #puppet-openstack
08:58 aldavud_away joined #puppet-openstack
09:11 mkoderer joined #puppet-openstack
09:22 Zordrak joined #puppet-openstack
09:22 Zordrak joined #puppet-openstack
09:28 dgollub joined #puppet-openstack
10:16 peterstac joined #puppet-openstack
10:36 sanjayu joined #puppet-openstack
11:10 openstackgerrit Lukas Bezdicka proposed a change to stackforge/puppet-nova: Don't manage the nova uid/gid if nova_{user,group}_id is undef  https://review.openstack.org/104583
11:15 gildub joined #puppet-openstack
11:20 dmsimard joined #puppet-openstack
11:31 tfz joined #puppet-openstack
11:40 tfz joined #puppet-openstack
11:57 dprince joined #puppet-openstack
12:09 sanjayu joined #puppet-openstack
12:18 sanjayu joined #puppet-openstack
12:26 sanjayu joined #puppet-openstack
12:29 tfhartmann joined #puppet-openstack
12:55 JayJ_ joined #puppet-openstack
13:11 dvorak joined #puppet-openstack
13:14 openstackgerrit Soren Hansen proposed a change to stackforge/puppet-ceilometer: Use 2.2 as the default mysql_module  https://review.openstack.org/107364
13:15 openstackgerrit Soren Hansen proposed a change to stackforge/puppet-cinder: Use 2.2 as the default mysql_module  https://review.openstack.org/107365
13:15 openstackgerrit Soren Hansen proposed a change to stackforge/puppet-glance: Use 2.2 as the default mysql_module  https://review.openstack.org/107366
13:16 openstackgerrit Soren Hansen proposed a change to stackforge/puppet-heat: Use 2.2 as the default mysql_module  https://review.openstack.org/107367
13:16 openstackgerrit Soren Hansen proposed a change to stackforge/puppet-keystone: Use 2.2 as the default mysql_module  https://review.openstack.org/107368
13:16 openstackgerrit Soren Hansen proposed a change to stackforge/puppet-neutron: Use 2.2 as the default mysql_module  https://review.openstack.org/107370
13:17 openstackgerrit Soren Hansen proposed a change to stackforge/puppet-nova: Use 2.2 as the default mysql_module  https://review.openstack.org/107372
13:17 openstackgerrit Soren Hansen proposed a change to stackforge/puppet-trove: Use 2.2 as the default mysql_module  https://review.openstack.org/107374
13:23 openstackgerrit Soren Hansen proposed a change to stackforge/puppet-ironic: Add support for MySQL module 2.2  https://review.openstack.org/106063
13:38 JayJ_ joined #puppet-openstack
13:43 sputnik1_ joined #puppet-openstack
13:44 mjblack joined #puppet-openstack
13:45 mjblack joined #puppet-openstack
13:51 pradk joined #puppet-openstack
14:07 imcsk8 joined #puppet-openstack
14:21 doo joined #puppet-openstack
14:31 doo_ joined #puppet-openstack
15:15 rstarmer joined #puppet-openstack
15:16 mdorman joined #puppet-openstack
15:22 daneyon_ joined #puppet-openstack
15:38 aldavud joined #puppet-openstack
15:56 StarBeast joined #puppet-openstack
15:57 daneyon joined #puppet-openstack
16:02 mgagne michchap: welcome back
16:07 daneyon joined #puppet-openstack
16:10 daneyon_ joined #puppet-openstack
16:12 thartmann joined #puppet-openstack
16:12 rmoe joined #puppet-openstack
16:23 aimon joined #puppet-openstack
16:27 sputnik1_ joined #puppet-openstack
16:27 sputnik1_ joined #puppet-openstack
16:28 sputnik1_ joined #puppet-openstack
16:28 StarBeast joined #puppet-openstack
16:36 rstarmer joined #puppet-openstack
17:01 tfz joined #puppet-openstack
17:09 mgagne hogepodge, EmilienM: Can you re-approve this one? https://review.openstack.org/#/c/105323/
17:09 mgagne hogepodge, EmilienM: I rebased against master
17:17 openstackgerrit joined #puppet-openstack
17:20 mmaglana joined #puppet-openstack
17:25 hogepodge joined #puppet-openstack
17:30 StarBeast joined #puppet-openstack
17:30 rstarmer joined #puppet-openstack
17:39 shakamunyi joined #puppet-openstack
17:40 shakamunyi joined #puppet-openstack
17:43 csschwe joined #puppet-openstack
17:47 arnaud_orange left #puppet-openstack
17:50 xarses joined #puppet-openstack
17:50 rstarmer joined #puppet-openstack
18:07 tesh joined #puppet-openstack
18:07 tesh تحذير
18:07 tesh warning
18:07 tesh you may be  watched
18:07 tesh do usa&israel use the internet(facebook,youtube,twitter, chat rooms ..ect)to spy??
18:07 tesh do usa&israel use the internet 2 collect informations,,can we call that spying??
18:07 tesh do they record&analyse everything we do on the internet,,can they harm you using these informations??
18:07 tesh warning
18:07 tesh joined #puppet-openstack
18:09 tesh joined #puppet-openstack
18:09 tesh joined #puppet-openstack
18:09 tesh left #puppet-openstack
18:10 EmilienM mgagne: done
18:10 mgagne EmilienM: ty sir
18:14 csschwe joined #puppet-openstack
18:29 hogepodge aimon mgagne michchap EmilienM: heads up, crinkle is starting to work out the aviator design work. we should be getting a wip spec up relatively soon.
18:30 EmilienM hogepodge: nice
18:30 mgagne hogepodge: any code yet?
18:30 mgagne hogepodge: =)
18:31 hogepodge we’re done, just waiting for payment. These open source projects never pay on time.
18:31 StarBeast joined #puppet-openstack
18:32 aimon thx hogepodge!
18:36 social joined #puppet-openstack
18:39 hogepodge aimon and mmaglana did we ever figure out how to strip down the requirements so that we can depend on preinstalled packages instead of gems?
18:39 aimon it no longer has any external requirements
18:39 aimon see module I created for it
18:39 hogepodge is the module a fork?
18:40 aimon of what?
18:40 hogepodge I saw it, and want to make sure we have a way to keep it in sync with the mainline aviator
18:40 aimon ah
18:40 aimon aviator insidfe it is not modified from mainline
18:40 hogepodge crinkle ^^
18:40 aimon mmagalan made mainline Aviator support older ruby and removed many external deps
18:40 hogepodge ok, so we just need to make sure we stay in sync
18:41 aimon yep
18:41 aimon *if there are features we need
18:41 aimon .. like when he does API V3
18:41 mmaglana hogepodge: yup. i just need to tell aimon (who's across the table from me) whenever i have a new release of aviator.
18:41 aldavud joined #puppet-openstack
18:41 hogepodge I know we talked about this before, but it’s fallen off the bottom of my stack
18:41 mmaglana :-)
18:41 aimon np hogepodge :)
18:41 hogepodge mmaglana we already want to add new api’s (like endpoints) (unless we missed the methods for those)
18:42 mmaglana hogepodge: cool. looking forward to the PRs at github!
18:42 mmaglana what endpoints btw?
18:43 hogepodge mmaglana keystone service-create
18:43 mmaglana cool.
18:44 hogepodge we’re thinking about how to pull the aviator library in to make it availabe to all of the modules, then doing piece by piece replacement of the calls in the provider
18:44 hogepodge One question being should be replace the current provider implementations or create a new provider
18:44 hogepodge Also, how to set up credentials so that we’re not depending on conf files and services running on localhost
18:44 aimon doing a new provider would allow a gradual switchover...
18:44 mmaglana hogepodge: re credentials, aviator accepts hashes now.
18:45 aimon might be more manageable that way
18:45 mmaglana hogepodge: see "Configure with an in-memory hash" at http://aviator.github.io/www/
18:45 hogepodge mmaglana aimon, right, but thinking about it from the point of view of how do you specify to puppet? Just have hashes in the manifest? Read an openrc file? Read the conf files? Some specification or ordering of that so you can fail over from one to another?
18:46 aimon hogepodge that is outside the scope of the module really.. and would be different for various use-cases
18:46 aimon i would personally put it in hiera
18:47 aimon others may use other methods
18:47 hogepodge right now puppet reads the conf files, and I think it’s sub-optimal
18:47 hogepodge but if the files exist, useful
18:47 aimon for the community OpenStack manifests.. probably just args liek all other auth info
18:48 aimon oh!
18:48 aimon yes that is not best idea
18:48 aimon I never like relying on potentially pollutred conf files
18:48 aimon *polluted
18:49 aimon will we be able to detect if a car is defined multiple times?
18:49 aimon which on is right?
18:49 aimon damn.. s/car/var/    s/on/one/
18:50 aimon having auth as args to modules is a more sane option
18:53 aldavud_ joined #puppet-openstack
18:53 hogepodge def. reuse variables. Feels like the auth source should be defineable. Like contiue to use conf files by default (then deprecate), but allow for explicit configuration
18:58 aimon +1 to that
19:06 mmaglana hogepodge: regarding aviator's repository. i spent some time working on it using openstack-gerrit workflow and it was too slow. reason is i test aviator in ruby 1.8, 1.9, 2.0, and 2.1 and i don't want to overload openstack infra with those jobs when i can use travis-ci for that. right now, i've reverted to my github workflow and the aviator repo is stackforge is not synced up with the latest. don't know if that affects you or not
19:06 mmaglana so just a heads up.
19:30 hogepodge mmaglana that’s good to know. I know that the infra team isn’t fond of not having stackforge be master, so it’s something to keep in mind.
19:31 mmaglana hogepodge: yeah, i'll need to ping them about how to resolve this. worst scenario is that i'll have to take it off stackforge.
19:32 hogepodge mmaglana that would be sad, but understandable. I don’t think infra is being overloaded, we test the modules on five or six different versions of puppet
19:37 mmaglana hogepodge: interesting. i'll look into it some more. one advantage of using openstack infra though is that i cand run the tests against a live openstack env. i have to rely on VCR-based mocks when using travis ci. something for my todo list.
19:41 StarBeast joined #puppet-openstack
19:54 dgollub joined #puppet-openstack
20:01 rstarmer joined #puppet-openstack
20:10 aldavud joined #puppet-openstack
20:20 bodepd mmaglana: It's a little early for it to be used, but I've been working on a way to install just enough bits of openstack-infra to gate upstream
20:21 bodepd mmaglana: right now my stuff relies on heat and requires an openstack endpoint to run against
20:21 bodepd that being said. I just hooked it up to monitor puppet-* changes
20:22 bodepd I don;t think it is going to create comments, and I'll be watching
20:22 bodepd someone just let me know if it starts spamming
20:22 mmaglana bodepd: sounds cool!
20:23 mmaglana bodepd: where are you hosting your instance/environment?
20:23 bodepd rackspace cloud
20:23 bodepd https://github.com/bodepd/puppet-openstack-gater
20:23 bodepd I don't think it's usable outside of rackspace cloud yet
20:23 bodepd or for anyone else besides me
20:24 bodepd :)
20:24 bodepd but in a couple of weeks it should be pretty usable
20:25 mmaglana neat
20:26 bodepd git log -u
20:26 bodepd oops :)
20:28 mjblack joined #puppet-openstack
20:46 rstarmer joined #puppet-openstack
20:47 JayJ_ joined #puppet-openstack
20:52 rstarmer joined #puppet-openstack
20:57 bodepd now someone submit a patch ;)
20:58 aldavud joined #puppet-openstack
21:01 JayJ_ joined #puppet-openstack
21:03 aldavud hogepodge: ping
21:05 aldavud hogepodge: https://github.com/puppetlabs/puppetlabs-openstack/blob/master/manifests/profile/rabbitmq.pp#L16 configures rabbitmq as ram instead of disk node for me
21:06 JayJ_ joined #puppet-openstack
21:06 aldavud hogepodge: I tried experimenting with setting the hostname to the management address but didn't find a working config
21:07 hogepodge alvadud hm, ok
21:07 aldavud this is on ubuntu precise btw. not sure if this has any influence
21:07 hogepodge I want to update the underlying code. That method of configuring is deprecated.
21:08 hogepodge I’ve been testing on trusty now, haven’t had a lot of luck with precise
21:09 rstarmer joined #puppet-openstack
21:09 aldavud I think I had the same issue on trusty, but currently don't have a test setup to verify
21:10 hogepodge wouldn’t be surprising.
21:10 aldavud I will quickly reprovision to verify
21:11 hogepodge I’m seeing memory starvation on my test node, probably because of that issue
21:12 aldavud my main issue is that I lose all queues upon restart
21:13 aldavud which wasn't an issue, as puppet could always recreate them
21:13 aldavud but it breaks the upgrade to icehouse
21:15 hogepodge ok, looking at code and thinking
21:16 aldavud so if I replaced the mgmt address with the hostname instead it works for me: https://github.com/ninech/puppetlabs-openstack/blob/master/manifests/profile/rabbitmq.pp#L25
21:17 hogepodge What’s strange is that it should default to disc
21:18 aldavud I think rabbit gets confused between rabbit@control and rabbit@172.16.33.4 and then thinks there's two instances and the other one is the disk one
21:19 hogepodge I’m spinning up a trusty instance right now. Let me check. If it works out fine I can get a new release out soon.
21:20 hogepodge The problem is I’m using that as a placeholder for when I start adding HA to the module, so ::hostname won’t be robust in the long term.
21:20 aldavud true
21:24 hogepodge from a 2011 post, “Right now, you can't really use IPs to refer to nodes.  We've got a bug open for that, but nobody's working on it, so it's unlikely to get fixed soon.”
21:25 hogepodge So I think that the solution is to use hostname, then figure out how to configure with hostnames when I build out HA.
21:25 hogepodge Good catch, thanks aldavud. Rabbit is definitely unhappy with ips, which I think is silly, but I usually think things that don’t work the way I want them to are silly (which is often silly
21:26 aldavud :D
21:28 hogepodge I’m also wondering about the rest of my configs. This may be a bigger bug that I need to investigate.
21:29 aldavud IPs vs hostnames you mean?
21:31 JayJ_ joined #puppet-openstack
21:31 aldavud so one other very random thing with rabbit is that it seems to 0.0.0.0 no matter what, but connects to itself via the given hostname
21:31 aldavud s/to 0.0.0.0/to bind to 0.0.0.0/
21:32 hogepodge it’s dumb. you have to have the hostname in the hosts file or through dhcp
21:32 aldavud yes
21:32 aldavud but as long as rabbit runs only on one host, you can even fix the hostname to 127.0.0.1 in the hosts file
21:34 aldavud and still connect remotely due to that behaviour
21:56 aldavud ok, I can't recreate with the master branch which uses the puppetlabs-rabbitmq 3.1.0 as the deprecated cluster_disk_nodes correctly sets "disk" in rabbitmq.config, this seems to be limited to stable/havana using 2.1.0: https://github.com/puppetlabs/puppetlabs-rabbitmq/blob/36f37c6c57728f4313ec678ae1fc6e3a2f758425/templates/rabbitmq.config
21:58 aldavud 3.1.0: https://github.com/puppetlabs/puppetlabs-rabbitmq/blob/25420cb3c26c9a8432ab519d43c094d614d4fec4/templates/rabbitmq.config.erb#L6
22:02 hogepodge yeah, I upgraded
22:02 michchap mgagne: thanks :)
22:02 hogepodge rabbit mq is broken in mine, though.
22:02 hogepodge I’ll work on a fix for 4.1.1
22:05 aldavud great!
22:05 aldavud ok, need to get some sleep!
22:17 rstarmer joined #puppet-openstack
22:53 gildub joined #puppet-openstack
22:55 rstarmer joined #puppet-openstack
22:58 openstackgerrit A change was merged to stackforge/puppet-neutron: Introduce neutron:config to manage neutron configuration  https://review.openstack.org/97281
23:04 aimon joined #puppet-openstack
23:04 mgagne hmmm looks like the kind of change that should be documented somewhere: https://github.com/stackforge/puppet-ceilometer/commit/bea61ade60e87375c44d242ae196292776332d6c
23:04 mgagne was wondering why those configs where purged from nova.conf
23:11 xarses mgagne: https://bugs.launchpad.net/puppet-neutron/+bug/1343009 << 4 days of my life
23:12 xarses :/
23:12 mgagne xarses: that's why I disabled notifications =)
23:12 mgagne xarses: and I got errors related to rubygems import too for that one
23:13 xarses oh, I also have to break the ruby resource providers calls to keystone
23:13 xarses otherwise the response is cached with no information
23:13 xarses raises some ugly nil error
23:13 mgagne xarses: didn't we fix it?
23:13 xarses not quite sure, using new neutron with older keystone
23:14 mgagne xarses: unless I'm running a fork of puppet-keystone...
23:14 xarses but call that it uses looks the same
23:14 mgagne xarses: yep https://github.com/mgagne/puppet-keystone/commit/049aed5bbb3473c49ab74ccf31534a9ba366f189
23:14 xarses https://review.openstack.org/#/c/103280/29/deployment/puppet/keystone/lib/puppet/provider/keystone_tenant/keystone.rb
23:15 xarses L:89 raises issues, so dropped ||= on L:29
23:16 xarses yep, you're dropping some of the ||= too
23:17 mgagne caching is a nightmare. if you happen to create resources based on others you JUST created, you are screwed
23:18 mgagne xarses: so notifications configs have to be located along neutron-server right?
23:19 mgagne xarses: and not word or trace about neutron-server in neutron::server::notifications, super awesome
23:23 xarses mgagne: i added 'Nova_admin_tenant_id_setter<||> ~> Service['neutron-server']'
23:23 xarses to server.pp
23:23 xarses and it takes care of it
23:23 xarses maybe it should be the other way, not sure yet
23:24 xarses also, db-sync dosen't allways run before all the config either
23:24 xarses but thats a seperate issue
23:24 mgagne xarses: looks ok to me, I guess the issue is that nova_admin_tenant_id_setter is not a neutron_config which already trigger a refresh of neutron-server on change
23:24 mgagne xarses: don't start me on db-sync...
23:25 mgagne xarses: in my hiera:
23:25 mgagne # NOTE(mgagne) Disable db_sync, it's the devil itself!
23:25 mgagne neutron::server::db_sync: false
23:25 xarses lol
23:25 xarses yes, but in some versions neutron-server forgets to stamp the db
23:26 xarses so it explodes randomly
23:26 mgagne xarses: yep, that's why I disabled it as I was running Havana until now and db wasn't stamped
23:27 mgagne xarses: fun, now policy syntax in Icehouse for Keystone :D
23:27 mgagne new*
23:28 * xarses mumbles
23:32 sputnik1_ joined #puppet-openstack
23:37 openstackgerrit Andrew Woodward proposed a change to stackforge/puppet-neutron: Make Nova_admin_tenant_id_setter notify neutron-server.  https://review.openstack.org/107529
23:39 doo joined #puppet-openstack
23:48 xarses mgagne: is there a test pattern for that relation ^
23:50 mgagne xarses: it's those weird tests difficult to write
23:52 mgagne xarses: first, test should be against neutron::server::notifications as it's the only one using nova_admin_tenant_id_setter. I guess you could test that the relation between nova_admin_tenant_id_setter['nova_admin_tenant_id'] and Server['neutron-server'] is properly done.
23:53 xarses hmm, then the relation would need to be declared in neutron::server::notification
23:53 xarses otherwise the test will have to evaluate both classes
23:54 xarses hmm, no it would also end up not matched
23:55 xarses since neither evaluates both
23:55 mgagne xarses: yep
23:55 xarses really, looking at notification, i don't like that its seperate
23:55 mgagne xarses: but you have to include neutron::server to use neutron::server::notifications no?
23:56 xarses no
23:56 mgagne xarses: I don't have a strong opinion about it except that it could be used as an example of how to split configs in multiple classes
23:56 xarses i think that notification should be moved into server
23:57 xarses I don't know why its seperate anyway
23:57 mgagne xarses: like the plethora of configs related to messaging or database
23:58 mgagne xarses: as long as you can disable it, I'm fine with it in neutron::server
23:59 xarses I think the default config ends up enabling notify_nova_on_port_status_changes anyway
23:59 mgagne xarses: but without nova authentication info
23:59 xarses which is why i started configuring it because i found auth failures with the call back

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary